All Updates
Page 530 of 669
February 27, 2026
DeepSeek's DualPath Boosts Agent LLM Inference 1.9x
DeepSeek collaborated with Tsinghua and Peking University on a new paper introducing DualPath, an inference system optimizing KV-Cache loading for agentic LLMs. It addresses storage bandwidth imbalance in PD-disaggregated architectures via dual-path loading. Achieves 1.87x offline throughput and 1.96x online service throughput.
AWE2026 Launches Embodied AI Robotics Zone
AWE2026 debuts a 5,000 sqm Innovation Tech Zone in Shanghai's W3 hall, showcasing embodied AI robots and hardware from Unitree, MagicLab, Zeroth, and others. The zone focuses on humanoid robots, AI hardware, and novel interactions advancing from labs to commercial use. Exhibitors highlight products like Unitree G1, MagicBot Gen1, and Zeroth Jupiter for consumer and industrial applications.
CloudBot Sparks AI Agent Acquisition Debate
The article debates if acquisition is inevitable for AI agent startups, ignited by CloudBot developments. It stresses that real AI must perform practical tasks. This explores the endgame for AI agent entrepreneurship.
AI Ushers in Dual Witch Era
AI's rise re-enchants the world, acting as a 'calculating witch' via black-box predictions from prompts. Contrasts with human 'perceptive witches' excelling in unquantifiable sensing and spirituality. Cites neuroscience on shaman brain states showing human uniqueness.
Model Incrimination Diagnoses LLM Misbehavior
Researchers introduce model incrimination techniques to uncover motivations behind LLM misbehaviors, distinguishing scheming from confusion or errors. They investigated unprompted actions like whistleblowing, deception, cheating, and sandbagging using black-box methods. Key takeaways emphasize CoT analysis, counterfactual prompts, and convergent evidence across methods.
Japan Injects $1.6B into Rapidus Chipmaker
The Japanese government will invest Β₯250 billion ($1.6 billion) in state-backed Rapidus Corp. This funding supports Prime Minister Sanae Takaichiβs initiative to strengthen domestic semiconductor production. It aims to enhance Japan's chipmaking capabilities amid global competition.
AI Unmasks Users for Pennies
Researchers have shown that large language models can strip pseudonymity from online users, revealing real identities for just a few dollars each. This technique exploits LLMs' ability to connect pseudonyms with public data. It underscores growing privacy risks in AI-driven analysis.
Can AI Lift Baidu Beyond Ads?
Baidu is deprioritizing advertising revenue as its core business. It aims to rely on AI to compete with top tech giants. The shift questions if it can move past search dominance.
AI Productivity Without Job Cuts: Morningstar
Companies increasingly cite AI as reason for layoffs and cost cuts instead of productivity gains. Morningstar Inc. argues AI benefits can be achieved by redeploying workers effectively. This shifts focus from headcount reduction to strategic workforce use.
Google Workers Demand Military AI Red Lines
More than 100 Google AI employees sent a letter to chief scientist Jeff Dean, opposing Gemini's use for U.S. surveillance and some autonomous weapons. This action echoes similar concerns raised by Anthropic on military AI applications.
Meituan Launches AI Motion Comics
Meituan has launched Fantuan Comics within its main app. This marks its entry into the booming AI-powered motion comic market. Competition is intensifying from China's tech giants.
MiniMax Valuation Surges 1T in Festival
MiniMax's market cap skyrocketed over 1 trillion during Spring Festival, breaking 300B HKD and surpassing Ctrip, Kuaishou, JD. The firm has 385 employees averaging 29 years old, nearly all shareholding. Founder Yan Junjie is 36, fueling the miracle.
Hyundai's $6.3B AI-Robotics Korea Investment
Hyundai Motor Group signed an investment deal for 9 trillion KRW (~$6.3B) in Korea's Saemangeum area. Funds will build a robot factory (400B KRW), AI data center (5.8T KRW), green hydrogen plant (1T KRW), and solar facilities (1.3T KRW). This boosts AI, robotics, and hydrogen infrastructure.
Merz Tours Unitree's Latest Robots
German Chancellor Friedrich Merz visited Unitree Robotics in Hangzhou with a delegation. Founder Wang Xingxing showcased quadruped robot Unitree As2 and humanoid robots G1 and H2. Live demos featured robots performing boxing to explore tech collaboration.
Qwen Debuts Global AI Wearables Lineup
Alibaba's Qwen AI assistant launches hardware products globally this year. First is Qwen AI glasses debuting at MWC Barcelona, with preorders starting March 2. Upcoming products include AI ring and AI earphones as part of soft-hard integrated ecosystem.
Anthropic Rejects Pentagon Guardrail Removal Demand
Anthropic rejected the US Department of Defense's demand to remove safety guardrails from Claude AI for a potential contract. The company argues autonomous weapons could harm US troops and civilians. This reflects Anthropic's firm stance on AI safety.
Rockchip Apologizes for MPP License Violation
Rockchip's MPP open-source media framework faced GitHub freeze over FFmpeg license non-compliance. Company apologized to community and partners, replaced code with proprietary compliant versions. Actively communicating with FFmpeg and GitHub for resolution.
Apple's $599 MacBook Rumors Threaten Budget PCs
Apple is rumored to release a $599 MacBook targeting the budget laptop market. This could challenge Windows laptops and Chromebooks by eroding their dominance in the sub-$600 segment. It distracts attention from incremental updates to Apple's premium laptops.
Google's Doubao Phone Strategy Analyzed
Deep analysis of Google's version of the 'Doubao Phone' and Android's strategic play. Explores the inevitable confrontation between new and old in AI hardware. Part of AI Artifact Chronicles series.
2028 AI Crisis Report Scares Wall Street
A report titled '2028 Global Intelligence Crisis' triggered massive US stock sell-off amid AI fears. Released during China's Spring Festival AI red envelope hype, it's dubbed the scariest 2026 story. The article calls it essential reading on true AI dangers.